numolus/app/frontend/io/xlsx.ts

/**
 * .xlsx import/export via SheetJS (the *only* place a spreadsheet library is
 * used, and never for formula evaluation -- that stays in our engine).
 *
 *   export: one worksheet per table. Formula cells are written as real Excel
 *           formulas when they translate cleanly (`;` -> `,`, `AVG` ->
 *           `AVERAGE`), always accompanied by the cached value from our engine.
 *   import: one table per worksheet. `cell.f` becomes a formula when our parser
 *           accepts it, otherwise the cached value is taken.
 */

import type * as XLSXNS from 'xlsx';
import { parseFormula } from '../engine';
import type { DocumentStore } from '../store/document';
import type { Sheet, TableObject } from '../store/types';
import { defaultTableView } from '../store/types';
import { rawString } from '../store/format';

/** SheetJS is the only spreadsheet library and it is loaded on demand. */
let xlsxModule: typeof XLSXNS | null = null;
async function loadXLSX(): Promise<typeof XLSXNS> {
  if (!xlsxModule) xlsxModule = await import('xlsx');
  return xlsxModule;
}

/** Swap top-level separators while respecting "..." and '...' quoting. */
function swapOutsideQuotes(src: string, from: string, to: string): string {
  let out = '';
  let q: '"' | "'" | null = null;
  for (const ch of src) {
    if (q) {
      out += ch;
      if (ch === q) q = null;
    } else if (ch === '"' || ch === "'") {
      q = ch;
      out += ch;
    } else if (ch === from) {
      out += to;
    } else {
      out += ch;
    }
  }
  return out;
}

function toExcelFormula(input: string): string | null {
  // input starts with '='
  let body = input.slice(1);
  body = swapOutsideQuotes(body, ';', ',');
  body = body.replace(/\bAVG\s*\(/gi, 'AVERAGE(');
  return body;
}

function fromExcelFormula(f: string): string | null {
  const body = swapOutsideQuotes(f, ',', ';');
  const candidate = '=' + body;
  try {
    parseFormula(candidate);
    return candidate;
  } catch {
    return null;
  }
}

function sanitizeSheetName(name: string, taken: Set<string>): string {
  const clean = name.replace(/[[\]:*?/\\]/g, ' ').trim().slice(0, 31) || 'Tabelle';
  let candidate = clean;
  let i = 2;
  while (taken.has(candidate.toLowerCase())) {
    candidate = `${clean.slice(0, 28)} ${i++}`;
  }
  taken.add(candidate.toLowerCase());
  return candidate;
}

// ---- export -----------------------------------------------------------

export async function documentToXlsx(store: DocumentStore): Promise<Blob> {
  const XLSX = await loadXLSX();
  const wb = XLSX.utils.book_new();
  const taken = new Set<string>();

  // one worksheet per table object, across all sheets
  const tables = store.doc.sheets.flatMap((s) =>
    s.objects.filter((o): o is TableObject => o.kind === 'table'),
  );
  for (const table of tables) {
    const ws: XLSXNS.WorkSheet = {};
    let maxC = 0;
    let maxR = 0;

    for (const key of Object.keys(table.cells)) {
      const [c, r] = key.split(',').map(Number);
      const raw = table.cells[key];
      const addr = XLSX.utils.encode_cell({ c, r });
      const value = store.cellValue(table.id, c, r);

      if (raw.startsWith('=')) {
        const f = toExcelFormula(raw);
        const num = typeof value === 'number' ? value : undefined;
        ws[addr] =
          num !== undefined
            ? { t: 'n', v: num, f: f ?? undefined }
            : { t: 's', v: rawString(value), f: f ?? undefined };
      } else if (typeof value === 'number') {
        ws[addr] = { t: 'n', v: value };
      } else {
        ws[addr] = { t: 's', v: raw };
      }
      if (c > maxC) maxC = c;
      if (r > maxR) maxR = r;
    }

    ws['!ref'] = XLSX.utils.encode_range({ s: { c: 0, r: 0 }, e: { c: maxC, r: maxR } });
    XLSX.utils.book_append_sheet(wb, ws, sanitizeSheetName(table.name, taken));
  }

  const out = XLSX.write(wb, { type: 'array', bookType: 'xlsx' }) as ArrayBuffer;
  return new Blob([out], {
    type: 'application/vnd.openxmlformats-officedocument.spreadsheetml.sheet',
  });
}

// ---- import -----------------------------------------------------------

export interface ImportedSheet {
  name: string;
  rows: number;
  cols: number;
  cells: Record<string, string>;
}

export async function xlsxToSheets(buffer: ArrayBuffer): Promise<ImportedSheet[]> {
  const XLSX = await loadXLSX();
  const wb = XLSX.read(buffer, { type: 'array', cellFormula: true, cellNF: false });
  const result: ImportedSheet[] = [];

  for (const name of wb.SheetNames) {
    const ws = wb.Sheets[name];
    const cells: Record<string, string> = {};
    let rows = 1;
    let cols = 1;
    if (ws && ws['!ref']) {
      const range = XLSX.utils.decode_range(ws['!ref']);
      for (let r = range.s.r; r <= range.e.r; r++) {
        for (let c = range.s.c; c <= range.e.c; c++) {
          const cell = ws[XLSX.utils.encode_cell({ r, c })] as XLSXNS.CellObject | undefined;
          if (!cell) continue;
          let raw: string | null = null;
          if (cell.f) raw = fromExcelFormula(String(cell.f));
          if (raw == null) {
            if (cell.v == null) continue;
            raw = cell.t === 'n' ? String(cell.v) : String(cell.w ?? cell.v);
          }
          if (raw === '') continue;
          cells[`${c},${r}`] = raw;
          rows = Math.max(rows, r + 1);
          cols = Math.max(cols, c + 1);
        }
      }
    }
    result.push({ name, rows: Math.max(rows, 20), cols: Math.max(cols, 8), cells });
  }
  return result;
}

/** Wrap an imported worksheet as a canvas sheet holding one table object. */
export function importedToSheet(imported: ImportedSheet, idx: number): Sheet {
  const stamp = Date.now().toString(36);
  const table: TableObject = {
    kind: 'table',
    id: `xlsx-tbl-${stamp}-${idx}`,
    name: imported.name,
    x: 40,
    y: 40,
    rows: imported.rows,
    cols: imported.cols,
    colWidths: {},
    cells: imported.cells,
    view: defaultTableView(),
  };
  return { id: `xlsx-sheet-${stamp}-${idx}`, name: imported.name, objects: [table] };
}